# VNCoreMLRequest An image-analysis request that uses a Core ML model to process images. ``` class VNCoreMLRequest ``` ## Overview The results array of a Core ML-based image analysis request contains a different observation type, depending on the kind of object you use: - If the model predicts a single feature, the model’s object has a non-`nil` value for and Vision treats the model as a classifier. The results are [`VNClassificationObservation`](/documentation/Vision/VNClassificationObservation) objects. - If the model’s outputs include at least one output with a feature type of , Vision treats that model as an image-to-image model. The results are [`VNPixelBufferObservation`](/documentation/Vision/VNPixelBufferObservation) objects. - Otherwise, Vision treats the model as a general predictor model. The results are [`VNCoreMLFeatureValueObservation`](/documentation/Vision/VNCoreMLFeatureValueObservation) objects. > Note: > Vision forwards all ``doc://Vision/documentation/Vision/VNObservation/confidence`` values from Core ML models as-is and doesn’t normalize them to `[0, 1]`. ## Topics ### Initializing with a Core ML Model [`init(model:)`](/documentation/Vision/VNCoreMLRequest/init(model:)) Creates a model container to use with an image analysis request based on the model you provide. [`init(model:completionHandler:)`](/documentation/Vision/VNCoreMLRequest/init(model:completionHandler:)) Creates a model container to use with an image analysis request based on the model you provide, with an optional completion handler. [`model`](/documentation/Vision/VNCoreMLRequest/model) The model to base the image analysis request on. [`VNCoreMLModel`](/documentation/Vision/VNCoreMLModel) A container for the model to use with Vision requests. ### Configuring Image Options [`imageCropAndScaleOption`](/documentation/Vision/VNCoreMLRequest/imageCropAndScaleOption) An optional setting that tells the Vision algorithm how to scale an input image. [`VNImageCropAndScaleOption`](/documentation/Vision/VNImageCropAndScaleOption) Options that define how Vision crops and scales an input-image. ### Identifying Request Revisions [`VNCoreMLRequestRevision1`](/documentation/Vision/VNCoreMLRequestRevision1) A constant for specifying revision 1 of a Core ML request. --- Copyright © 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)